Privacy Protection of Healthcare Data over Social Networks Using Machine Learning Algorithms.

Area of Science:

• Computer Science
• Information Security
• Healthcare Informatics

Background:

• Mobile medical care raises concerns about personal medical data privacy leakage.
• Existing k-anonymity and l-diversity models have limitations in fine-grained privacy protection.

Purpose of the Study:

• To propose a classified personalized entropy l-diversity privacy protection model for fine-grained user privacy.
• To address the issue of standard information entropy l-diversity models failing to differentiate between strong and weak sensitive features.

Main Methods:

• Developed a customized information entropy l-diversity model.
• Distinguished between solid and weak sensitive attribute values to improve attribute constraints.
• Reduced sensitive information to lower the probability of vital information leakage.

Main Results:

• Experimental results demonstrate minimized execution time and improved data accuracy.
• The proposed method enhances data accuracy and service quality compared to existing solutions.
• Reduced sensitive data and lowered the chance of crucial data leakage, enhancing healthcare data exchange security.

Conclusions:

• The customized information entropy l-diversity model effectively protects user privacy in a fine-grained manner.
• The approach enhances data accuracy while minimizing algorithm execution time.
• This method offers a more effective solution for secure medical data sharing in mobile healthcare.